feat(wren-ai-service): evaluation dataset curation app

add eval dataset curation app v1
- users can create new eval dataset
- Upload an MDL json file first
- Get question-sql-pairs given by LLM or you manually enter question and corresponding sql
- Do validation on each group of question, context and SQL, and move it to the candidate dataset if you think it's valid
- Save the candidate dataset by clicking the "Save as Evaluation Dataset" button.
- users can modify existed eval dataset by uploading the dataset
- Upload an MDL json file first
- Upload the evaluation dataset(.toml file) you want to modify, and please make sure the model in the dataset is the same as the deployed model
- Modify the evaluation dataset the same as you create a new one
- Save the candidate dataset by clicking the "Save as Evaluation Dataset" button.
- users can choose which OpenAI llm model to use
- users can type custom instructions for llm to generate question-sql-pairs
- users can preview data if sql is valid
update setup instruction on readme
- go to the eval folder and run poetry install to install the dependencies
- execute commands in the wren-ai-service folder
- to run the evaluation dataset curation app: make data_curation_app

Canner / WrenAI